KiOR: The inside true story of a company gone wrong. Part 3, “You’ve Cooked the Books”

Cherry-picking the data

In February 2011, John Hacskaylo published an internal R&D Update Report at KiOR, which now can be revealed. In it, he showed dramatic improvement in bio-oil yields produced at the Demonstration Unit. Allowing for a responsible level of oxygen content, sufficiently low for the product to be upgraded to a finished fuel, the data showed a 67 per gallon output. Hacskaylo extrapolated to 84 gallons per ton as a target for the future.

But there was a problem.

According to KiOR staff members of that time, the data was cherry-picked; the most optimistic data was being used. Favorable bursts based on a handful of moments in time.

In fact, when the Demo plant was lined up and operating in steady-state condition at least for a few days, what were described by KiOR staffers familiar with the data as the actual “measurable, recoverable and reproducible bio-oil yields with oxygen content of less than 15%” (the maximum for a refinery to handle), were in the range of 34-38 gallons per ton.  Substantially much lower that the bio-oil yields that Hacskaylo was reporting.

But it was worse than that.  The yields were lower than the Bio oil yields produced with same catalyst and process conditions at KiOR Pilot plant.

It was a dagger in KiOR’s path to commercial viability — because yields were supposed to go up, in the demonstration, which was intended to be an improved design based on pilot plant data. Not down. Not with the same process conditions and catalyst. Something was very wrong with the demo design. Something that was concerning Stamires and Charlie Zhang who were monitoring and analyzing the raw DEMO data and comparing them to those obtained from the KCR Pilot plant . But no one on the commercial management team would know about it for quite some time.

Worse, if this trend-line was valid, and could be extrapolated to larger size plants, then it was possible that Bio oil yields at the commercial-scale Columbus plant, when operational and using the same catalyst and process variables would be even smaller than the Demo.

The drag from pilot yields to demo yields was around 25%, so it was possible that commercial-scale yields would drop into the 20s. Well below financial viability under any foreseeable combination of circumstances.